bioinformatics, genomes, biology etc. "I don't mean to sound angry and cynical, but I am, so that's how it comes across"

A guide for the lonely bioinformatician

This may be a uniquely UK centric blog post but I suspect not.  Let me start with a brief story.  Sat with a coffee in our canteen a few weeks ago, I overheard a conversation between a few PIs about a grant application.  “Don’t worry”, the lead PI said,  “we’ve put money on the application to fund a bioinformatician”.  Good planning I hear you say, and I agree;  however,  note that none of the PIs in that discussion were themselves bioinformaticians; none of them can code;  put them in front of a Linux terminal and they wouldn’t know what to do.

Yes – we were witnessing the birth of yet another “pet bioinformatician”.  What I mean by this term is a single bioinformatician employed within laboratory based group. These guys are becoming more and more common in UK academic groups,  and it concerns me because it is possible they will become isolated and pick up bad practices as they don’t have a senior bioinformatician to guide them. It also concerns me that their career and profesional development might suffer.

Consider the opposite situation – how many bioinformatician PIs manage lab staff?  How could we possibly guide a young post doc on how to run gels, PCRs etc nevermind more complicated laboratory SOPs? We couldn’t – so why do lab-based PIs assume they can guide bioinformaticians?

Consideration has to be given to how you can develop and nurture a young or inexperienced member of staff when you have NONE of the skills they will need to develop to survive in their chosen field. How can you help them when you dont know yourself?

This guide is aimed at pet bioinformaticians, and is meant to guide them towards better career development.

1. Make friends with local bioinformatics groups
You may not be in the local bioinformatics group, but if there is one, seek them out, introduce yourself and make friends. Ask if you can attend their lab meetings and journal club. Tell the group leader you want to make sure you learn good practice in bioinformatics, and would like their help. If there isn’t one in your institute, where is the nearest one geographically? Can you travel to meet them? If so, do it; if not, attempt to skype into meetings etc. Develop electronic relationships with people and groups on the Internet. Develop a support group who will be able to help you with the kind of problems your lab-based group cannot.

2. Talk to your computing group
Find them, tell them what your work is about, what resource you will need and ask them how best to get access to those resources in the environment you exist in. If you don’t know what resources you will need, see point (1) above. Your local IT team will be essential to your sucess. Befriend the linux sys admin, they might save your life.

3. Obtain clear expectations
Speak to your manager and get them to outline exactly what their expectations are of you. If you are funded on a grant, get the grant application and read exactly what your manager has promised you will deliver. Prepare your manager for the possibility that their expectations may need to be altered, especially if they are unrealistic.

4. Rewrite your job description
Armed with (3) and with the use of (1) and (2), rewrite your job description, make your manager fully aware what you can deliver and what you can’t. Make them aware of how long things take, as they may not know. Do this as early as you can. If they disagree with your estimates of what you can deliver, ask for the support from (1). Give realistic estimations. Ask your manager to prioritise the objectives.

5. Papers
You need first author papers, just like any other scientist. Middle author papers will only get you so far. Be up front about this, ask your manager where your next first author paper is coming from. Explain that, for certain projects you will have done more work than the lab guys, so deserve first authorship (only do this if its true). If no possibility exists, ask to be allowed to develop your own ideas and publish those. Talk to the guys in (1).

6. Attend bioinformatics meetings
Your manager will want you to go to meetings relevant to the group’s research. Go to these, but also ask to attend bioinformatics meetings and workshops.

7. Try first, ask later
This is a delicate balancing act. Nothing will teach you better than just getting hold of some data or code and just giving it a go. Try. Sit and read and try as hard as you can to solve whatever problems you encounter. But be aware the solutions you come up with may be sub optimal. After a certain time period, ask for help. Show what you’ve done to those with more experience and ask for feedback. Take the best of what you have learned and any feedback you gained, and leave the rest behind. Noone likes the person who asks for help too early and expects someone else to tell them what to do; we all love a trier. But don’t take it too far; try your own way first, but at some point take a break and ask for feedback and assistance.

I’ve seen more than one project where the results were almost 100% crap because a bioinformatician acted in isolation and didn’t ask for help. Don’t be that person. Don’t let yourself be. Being around other bioinformaticians should mitigate this risk, as they will be able to spot where you are going wrong before it is too late.

Make sure you are valued as highly as other members of your group, that your career is nurtured, that your skills are developed in the same way that other members of your group enjoy. Make sure you work to clear and realistic expectations, and take an active role in (re)defining your role and job description.

If you’re in academia, publish. Please publish. Mid author is ok, but first author is essential if you want to be seen as anything other than a facilitator. Aim for two papers a year. If your manager can’t deliver that for you, do it yourself. Publish a review. Publish a comparison of software tools. Take the problem you encounter most, solve it and publish the solution. It will seem daunting at first, but ask for help. You can do it.

Look after yourselves, pet bioinformaticians 🙂

Update – 24/4/2013

Wow, did this post touch a nerve – definitely my most popular post in terms of retweets and first day visitors.  There are clearly a lot of pet bioinformaticians out there!

A few things.  Firstly, if you are the PI of a pet bioinformatician, there is no explicit or implied criticism of you here.  There is nothing wrong with you employing a bioinformatician in your lab.  Just look after them, and recognise you can’t give them everything that they need.  You can give them a lot, just not everything.

Secondly, there is nothing wrong with being a pet bioinformatician – it can be a really stimulating role, and opens your eyes to lab-based science.  I am not criticizing the pets either, I just urge you to look after yourselves 🙂

Finally, this great comment from the tweetome:

I completely agree, if you can, you should be doing this.  Don’t be a passenger in lab meetings, suggest things that can be done, never forget you are a scientist too, you can propose hypotheses and how these may be tested 🙂  Your ideas are valid and you bring something to the party that no-one else in the room can.

Update – 30/4/2013

There is a possibility we could host a workshop in Edinburgh to look at best practices in “embedded bioinformatics”, involving some presentations, break out groups, and ultimately authoring a document/paper that highlights the pros and cons of the system.  This would either be free to attend, or would incur a very small cost to cover expenses (tea/coffee/lunch).  Would you come?


  1. I agree with you. This will become more and more an issue, as experimental labs get bigger grants for obtaining high-throughput data and they need people to analyze it. I’ve seen this a lot around here. My suggested solution to this has always been to try to “embed” one of my students/postdocs in the other lab. The “embedded bioinformatician” implies that we have a collaboration between the labs, the bioinformatician is supervised by me, but he is right smack in the middle of the experimental frenzy.

    It hasn’t worked so far, mainly because it has been difficult to establish either formally or technically. Also, there is a risk for the embedded bioinformatician to be overwhelmed by numerous requests from students/postdocs or even the experimental PI without any clear objectives in mind, without the proper control. We always end up doing a standard collaboration, with a student in my lab analyzing data for another lab, besides his/her own project. I still think the “embedded bioinformatician” could work out.

    • Yes, I think the embedded model can work, but again, it is about expectations, about communicating needs and desires. The scientific need is to have the data crunched, but there are also the needs of the data cruncher, which shouldn’t be ignored.

  2. If you’re a pet bioinformatician and your manager thinks that all you need to do bioinformatics is a honking big gaming PC in the corner of the lab, then you definitely need to talk to your IT team. Your bioinformatics infrastructure needs to live in a data centre, either local or Amazon’s. If nothing else, it means you’ll be able to think without being deafened by fan noise.

    But then I would say that, being an IT guy…

    • Unfortunately, say “cloud” to a University IT guy and they shout “security risk” back at you….

      • If the IT guys can’t provide a local computing cluster or a private cloud then people are just going to take their chances on Amazon (which I noticed you doing recently). The IT bods should probably accept the inevitable and follow the users into the cloud to secure things.

      • Absolutely, I used AWS for a number of things, not least of which is that I can do whatever the hell I like to the OS (my local IT guys won’t let me do that on servers here).

        Whatever your setup though, for a bioinformatician, getting to know the IT guys is essential 🙂

      • I work contract for the military and it’s been a constant struggle with them shouting security risk at everything I and our programmer suggest in updating the computing infrastructure where I work…no clouds, no AWS, no bioinformatic software that isn’t approved (which takes 8 mos to a year), no open source software (you say ‘open source’; they hear ‘spyware, freeware, shareware’), all USB drives are locked (can’t use them to transfer data)…we are left doing our work on un-networked stand alone linux computers that they don’t like because none of them are Linux admins and therefore have very little ability to conduct oversight so 90% of the time I hear the word ‘no’.

        The IT guys are great, it’s the Information Assurance cadre that routinely block all our efforts and they tell the IT guys what they can and cannot do. Seriously, it’s like they think our programming and open source bioinformatic software is going to hack their system, become self-aware and start zapping people.

        We did succeed in getting a Linux cluster and finally got semi-approval to set up a separate research network with linux testing environment…who knows when that’ll go up.

        I spend a lot of time coddling and fighting ignorance than being the researcher I am supposed to be, a frustrating aspect of my job. They want to be technologically cutting edge yet are unwilling to test new software, try new set ups or entertain ideas of changing the current system.

  3. I am what you would term an embedded bioinformatician, although I usually think of myself as a cancer genetics researcher. It has been my experience that this position has a lot to offer if you are not purely interested in developing methods, but are more interested in applying them. There is a lot to say here, but I one point is that I recommend that anyone in this position make sincere efforts to learn some elementary wet lab procedures. Cultivate a friendly postdoc or grad student who will let you watch them and then try the methods on some replaceable samples. If you can do some simple procedures like a western blot, PCR, or taqman you will grow as a scientist. Your colleagues will stop giggling at the sight of the statistician wearing gloves after a few days. The benefits include: the equipment that was mysterious is no longer so strange. You can read papers more critically having done the work yourself. You will get some respect from your wet lab colleagues for having tried their work. You will gain a lot of understanding about why molecular biologists think and work the way they do. This is particularly important for collaborations, because people trained in algorithms and digital data (as I was) often do not understand why biologists seem to work with informed hunches and are comfortable going with an idea if there is a trend but not perfect statistical support. You will be able to use the language of molecular biology naturally after you’ve done some work. You will also gain a huge appreciation for how incredibly difficult it is to get any data at all, how often experiments fail for no obvious reason, where technical noise is introduced, and what kind of finding is more likely to reproduce. I am not saying that applied statisticians should become wet lab biologists, but that if you do just a little of this work a large world opens up to you. The same arguments apply in reverse to bench scientists, but this conversation is about the bioinformatics side.

    • Definitely, this is a great idea, but do you think the learning curve is too steep?

      • I don’t think it’s too steep if you indeed bring the background knowledge you should possess as a bioinformatician. I’ve done the same, and I agree with David. Doing some of the wet lab work yourself gives you a lot more insight on where data is coming from, and what the usual problems might be that specific protocols inflict on the data generated. I’ve found my biologist co-workers very happy to help me get some work done once they got over their giggling.

      • If you can bake, you can do all of the manual tasks required for straightforward lab procedures. The physical actions are not complicated. However, starting completely on your own is not going to work. Learning to do anything in the lab is an apprenticeship, where you stand behind someone and keep asking questions as you go. It’s different in kind from the smart but inexperienced programmer who would like to learn Python. That person can work his or her way through a tutorials, re-implement some complicated software in Python, and become competent without talking to an experienced coder. I see experienced post-docs go through the same process when they want to learn a new procedure; if no one in the lab knows how to set up ChIP-seq, you go find someone who has done it and learn from them. The learning curve really does kick in during the debugging process, because even routine lab work will fail to work as expected some of the time. That’s when you need experienced hands around. One crucial realization is the degree of trial and error required to get a result: expose the western longer, try a different antibody, try a different fixative, increase the RNA concentration, etc.

    • I have done both at one time. I started off running PCRs, AFLPs and sequencing reactions but then got interested in the informatics side of things. It is a great idea to have working knowledge of wet lab protocols but doing it yourself might not be the best way to go about it.

    • I work in a Bioinformatics Department at UNC Charlotte. One of the faculty (Jennifer Weller) developed a genomics lab class for our students to get practical, hands-on experience in lab techniques, including performing qPCR assays, creating libraries, and running a sequencer (Ion Torrent). Students from biology as well as computer science backgrounds take the class. Everyone learns a lot, but the c.s. students seem to benefit the most because they learn first-hand about the limitations of data and instrumentation. They also get to see how much a modern laboratory depends on computers to run key instruments.

    • Sarah (@SarahHCarl)

      13th September 2013 at 9:27 am

      What do people think about the inverse situation – a wet lab-trained biologist who learned to analyze her own data (in this case because there was no pet bioinformatician in the lab and no one around had done NGS before), discovered she really liked that part better than the wet lab part, and hopes to go work as a bioinformatician in the future, possibly in an embedded situation? I guess the disadvantage is a lack of formal training in computer science (though I have done some courses), but the advantage I see is that I’ve done the experimental techniques myself, I know how frustrating they can be and what the common pitfalls are, so I know what quirks and biases to expect in the data. At least that’s how I hope to sell it in the future!

      I recently attended a sequencing data analysis workshop, and the room seemed to be divided between the self-described “biologists” and “bioinformaticians.” Each group kept saying things like, “My biologist tells me this…” or “My bioinformatician says that…” I was sitting there thinking, I have those conversations inside my own head! 🙂

      • Hi Sarah…So I started as you did…primarily wet lab. During my Ph.D. (2003-2010) I was a molecular biologist, PCR (in all it’s forms), cloning, culturing of microbes, isolating phages, using the EM, radioactive probing and manipulating upwards of 500,000 clones for my study…hence the 7 year Ph.D. ya think?! Through that time I was primarily analyzing sanger sequence data and doing phylogenetic analysis. My advisor was not keen on my attending workshops early on in my Ph.D. (that’s another can of worms) and there was no ‘class’ persay on advanced phylogenetic analysis which I count in with bioinformatics and was using to answer speciation questions. So I hit the books and emailed collaborators and other mentors which had used the techniques and could ‘tutor’ me via email (slow going but it works). The great thing about the bioinformatics crowd, at least from my experience is how open they are to helping you un-f*ck your data! I found I liked it as well. 5 years into my Ph.D. I was adept at sequence analysis and phylogenetics, now NGS was getting under way and by the time I left it was the ‘next step’ along with everything else. I was finally able to start attending some workshops which were invaluable!!! I continued to get ‘tutoring’ from those who knew more. I combed manuals and was known for contacting developers and bothering programmers about formatting and syntax and I’ve broken more than one BEAST run by learning to manipulate and create XML. For NGS I attended workshops, read the literature, downloaded programs, read manual after manual (and I still can’t figure out the method behind the madness of f*cking 454 Roche output sometimes! Illumina is so much nicer sometimes to work with *romantic sigh*), took a programming class, forced myself into Linux, got accustomed to using Stackoverflow (though I NEVER post haha, just read) as well as python docs (docs.python.org). I am far from ‘pro-bioinformatician’ and I go back and forth from calling myself a bioinformatician and a microbial/viral ecologist…I feel like both. And the conversation in your head that you have? I have that too…on a semi-daily basis.

        My group wanted me for a post doc because of my bioinformatics not the viral ecology end (sniff sniff) though they thought it was a perk that I was all ‘viral and stuff’. And the bioinformatics end landed me my job, cutting my PostDoc short (which I have mixed feelings about). And now I was the line of part-time ‘pet bioionformatician’. So I’ve been on the computer primarily for the past 5 years (Ph.D/PostDoc/job). I am only now diving back into my own research/wet lab work now…and only at about 20%; 80% is still bioinformatician where I am at. And of course working on manuscripts 10%, continuing trainings where I can 10%…and yes I realize I’m up over 100% but I think the typical working scientists is usually working over the 100% metrics.

        My advice…if you are going to pitch yourself as a bioinformatician primarily, I think it’s a bonus you have a biology background (PIs will like that) you can understand and perhaps ‘speak’ two languages now. But I encourage you to develop your computational end, the informatics to your bio…I am hoping you are out of the GUI stage by now and can bash around a Linux pretty decently. If you haven’t started learning scripting or haven’t mastered an editor (I use VIM but I hear some like Emacs) I would focus on that. Eventually, if this remains your career path…start learning some basic computer science and a programming language. I picked Python (my husband is a python programmer/software engineer) and well…if his wife was going to program, damnit it was going to be python and she was going to learn VIM…I mentioned Emacs and his eyes about rolled out of his head. I also read through this book: http://www.amazon.com/Practical-Computing-Biologists-Steven-Haddock/dp/0878933913, it was a good start.

        And finally consider a certificate or master’s degree aimed at bioinformatics…University of Edinburgh has one but they are popping up everywhere to meet the training demands. Coursera has computer/programming classes available on the web as well as EdX (I took a python class through them).

        It amazes me sometimes how some in my field trivialize what software engineers and programmers can do. Having been on the learning end, I routinely bang my head on the computer and cuss through code and formatting. It’s taken years for some of my peers and advisors from the pre-NGS days to appreciate what computing does for them.Software engineers and programmers go to school for years, or have just been a ‘computer guy’ since birth practically, to do what they do and they think differently than we do as biologists. How we think is utter chaos to them :). The perk that you is that you’ve been apart of the chaos and can hopefully facilitate organization and understanding of ‘said chaos’ to those who don’t have a biological background. The flipside for is you now are learning a new language and new rules. Often times biologists want to learn in 2 weeks what takes 2 years to fully learn, appreciate and understand.

        Basically, be patient with yourself, never stop learning, asking questions and troubleshooting and you’ll be good to go. Best of luck to you!

        Whew, that was long, ha…sorry!

      • Sarah (@SarahHCarl)

        13th September 2013 at 2:53 pm

        Hi Mel,

        Wow, thanks for such a detailed reply! Good to hear how well it’s going for you, coming from a similar place as I am! I’m definitely sold on Linux now and I do a fair amount of Perl scripting (although I’ve been getting my feet wet with Python too 😉 ), as well as some playing around with R. Basically I’ve used what seems to be the prevailing tactic around here of “Try stuff, if it doesn’t work look it up on the internet and see if you can find another way to do it, then try that, etc…” Sometimes this is a bit redundant – such as when I spent about a month when I was just starting to learn perl writing a script to find overlapping entries in .bed files only to realize afterwards that BedTools exists – but you learn a lot!

        I’ve checked out some of the offerings on Coursera and found them fairly useful – I did a data science class recently that introduced me to relational databases and some Big Data stuff like MapReduce, so that was fun. I’m trying to keep expanding my computational repertoire through side projects as much as possible. And Stackoverflow is an awesome resource, as is SeqAnswers. For NGS, I started out using Galaxy and then decided that I wanted to know more about what was being done to my data under the hood, so I tried replicating some other people’s bash pipelines and eventually settled on a workflow that I’m pretty happy with, although of course it’s constantly being updated and tweaked.

        I like what you said about having been part of the chaos – that’s definitely true. I think it’s helpful to understand the practical aspects of experimental design and real-world limitations like cost, scarcity of starting material, etc. But even I am surprised now at how little some colleagues understand of what goes into processing the data that they generate (and even more surprising is the lack of interest some of them show in learning more about it). I guess that’s why the idea of the embedded bioinformatician appeals to me, as a position that is involved in projects from beginning to end and is still close to the biology.

        Cheers, and good luck with your research!

      • This is cool. I too am a biologist and we have bioinformaticians here at my institute but I’m hoping to become at least passably competent at running some of the basic analyses, at least shooting for reasonable confidence in my analyses for publication purposes. I resonate with the sentiment of trying to learn 2 years worth of stuff in 2 weeks. I feel like everything is piecemeal at the moment (I’m at the stage where I’m f*cking up linux, can hack and slash code and occasionally make it work with amazing amount of hairpulling). I have some basic background in ecological stats/modelling in R plus learnt some basic perl for a class project some time back. Certainly don’t know enough computing/math but that would probably take another lifetime. Also tapping into any local user group meetings sounds like a brilliant way to learn. Open to more ideas about how to build basic competency in bioinformatics for those of us coming at it from the other end. Maybe I’m not thinking about it right, but from that end seems quite disorganized to me. Thanks Sarah and Mel for sharing your experiences and for the resources. I’ll definitely check them out!

  4. Another “pet bioinformatician” here. My group works at a very high pace, and that keeps me usually serving their continuous requests (what constitutes a very interesting and fun activity by the way). I had a PhD in Machine Learning before I started working as a bioinformatician, but I had to learn a whole lot of new things. Nothing wrong there. I always loved to learn things. I know the way to go is 90% of what you are saying in this guide, and I would love to follow that path, but in my case, I have to add to the equation the fact that I am the father of three little children. That makes me A LOT less competitive. Anyway, I am trying my best, but at nights I am just too tired to try to devise a first-authorship.

    Maybe the key point in the guide is the first one. Association. Collaboration. I think that’s the way to go.

    • Indeed, and if you have a happy career in the job you have, and there is no risk to that, then there’s nothing wrong with the way you’re going. A word of caution though – if your job ever stops being secure, you may find yourself on the job market, and those missing “first authors” may count…

      • As someone who has chosen to be a lone statistician in a lab, albeit on my own fellowships, and who now has three small children, I understand. If you can’t get first author, ask for joint first. Increasingly, I think the best science will involve stats and wet lab scientists working hand in hand. If you were central to a project make sure you help write the whole paper (not just the stats and informatics bits), and talk to your boss about your need to build a publication record.

  5. liked the “embedded” idea and would be wonderful to try your hands at doing experiments. Another tip is to talk with other experiment groups who are doing high-throughput experiments. one never knows when a method developed a while ago for one task will be useful in a completely different scenario. the new found appreciation does make you feel good.

  6. Georgia (@gkapatai)

    24th April 2013 at 9:02 am

    I am just starting my new post as a ‘pet bioinformatician’ in a public health setting. Having developed mostly through a lot of self-learning and online courses I was looking forward to be a part of a group but the lure of a permanent job was too big to ignore. However, I now found myself in a position with no support. I think all the points mentioned in this guide are very helpful and I have already initiated communication with a few local bioinformatics groups and the computing group on site. What I find very difficult is finding a good bioinformatics workshop. There are quite a few happening this summer but I found it difficult to choose the right one for me.

    I found this: http://staff.icar.cnr.it/cannataro/bbc2013/ which I might be able to help me understand the infrastructure and data management aspect of my current job description.
    I also found this: http://www.nextgenseq.co.uk/ that provides hands-on training with NGS data. Having worked mostly at the end of the pipeline (creating scripts to analyse genome sequences) I desperately need to get a better understanding of the QC and assembly steps using actual data so that I can actually go into the pipeline and change it if necessary.
    There are also the EBI training courses.

    Any suggestions?

  7. Great post as usual. 🙂 I am definitely walking the line between ‘pet’ and scientist in my current lab. Year 1 was all ‘pet’ slash mop up person for all the stuff they’ve done the past year and had no one to fix/analyze. I am now in year 2, have my own funding for my own research and am hoping to start actually doing my own research and publishing some papers. I’m 3 years out of my Ph.D. and only have one first author pub at the moment, though I’m a middle author on several. I’m still doing some mop up on the side but I’m hoping the transition goes well.

    I am forming relationships with other computational biologists and computer scientists that have more background than myself and it’s super overwhelming at times. Just because you spent 7 years getting a Ph.D. doesn’t mean you are close to knowing all you need to know for your job/career. It’s frustrating, humbling and exciting.

    BTW I presented your ‘What it takes to be a bioinformatician’ at a recent training workshop I conducted in bioinformatic sequence analysis. I find there is still a general ignorance in the field of what a bioinformatician ‘can deliver’. Many still believe we are black boxes, that we push buttons, that we spit out golden ‘pretty picture’ analysis for the next big publication without much effort. It’s taken over a year to convince my boss, I’m not slacking, bioinformatic analysis takes time and investment…that it’s more complicated–just because I can do it on a computer doesn’t make it ‘easy’. I also have to convince collaborators that they need to adjust their wet lab protocols sometimes, that bioinformatics will not magically make suboptimal wet lab protocols and crap data publishable…junk in junk out. If you set up or run your assay/experiment incorrectly, bioinformatics can’t ‘save’ you…rerun the damn assay. You’d be amazed how many don’t ‘get’ that still.

  8. I am pet ‘bioinformatician’, or more precisely ‘computational biologist’. Started out knowing almost nothing, worked hard last 2 years, published some, but have become jaded and now considering career change.
    If I read this article two years ago I might have been way better off. Tryin to connect with a satellite group have its limitations, and what I needed was direct day to day contact.
    For those starting out, heed the points above, it’s essential!

  9. I reinforce the role of training, in what concerns “pet bioinformaticians” and their role out there.
    Advising users to train themselves (obtaining kills and relative autonomy) does not out-shadow the role of the “pet”. On the contrary, it allows her/him to concentrate on higher levels of user guidance, lab/institutional strategies, etc. “Pet”s can do a valuable service in advising users to pick the right training for their profile. resourcing to GOBLET ( http://mygoblet.org/ ) and to the iAnn based service available at its precursor website BTN (www.biotnet.org) can be effectively used to create awareness on training availability worldwide.
    I run the GTPB programme in Oeiras,PT
    and I can testify that trained users make much better work with “pet” bioinformaticians when the when they are there, and enjoy direct exploration of the data that they generate all the time.

  10. Great post! Thank you for calling attention to this important issue. I agree completely with all the above.

    I think it will be good to note here the experience with this matter in Switzerland: For several years now the concept of “embedded bioinformatician” has been implemented and even formalized by the Swiss Institute of Bioinformatics (SIB). Many e-bioinfo people (including myself) are members of wet labs and also affiliated with a bioinfo group, which is a member of the SIB. This set up allows for exactly what is called for by the most important points above (especially number 1).

    I’m an e-bioinformatician in a group studying evolutionary genomics of ants, and I’m affiliated with a computational molecular evolutionary group. I attend their group meetings and I benefit from their experience and the advise of a computational PI. E-bioinfo are given membership in the SIB and we attend SIB conferences and workshops. So there’s a lot of interactions developing among e-bioinfo in neighboring groups and with the people in the purely computational groups. There is also great IT infrastructure centrally managed in Lausanne by the Vital-IT team, which serves both bioinfo groups and e-bioinfo in wet labs. This set up works very well for those otherwise lonely bioinformaticians. So I would recommend that bioinfo leaders elsewhere take a look into the Swiss experience (talk to the people who led this development, especially Ioannis Xenarios) and consider implementing something similar in your own country. I know some of the UK community, where a lot of great bioinfo is done, so I would think such an initiative (by EBI?) would do good to address the needs raised in this discussion.


  11. Mick, I couldn’t agree more. I’ve actually seen worse than this: Biology PIs taking on Bioinformatics PhD students.

    My view: research councils should not fund these grants unless there is a Bioinformatics co-I on the grant. If I were refereeing such a grant, I wouldn’t hesitate to recommend it be rejected on these grounds. It has to become the norm that research staff on the grant must be supported by appropriate expertise among the investigators and their groups.

    Also, I think it is critical that universities/research institutions set up bioinformatics support services. That is, groups of people to carry out ‘routine’ bioinformatics tasks in support of other research projects. These groups are best populated by people who are not aspiring to an independent research career (traditional postdocs), but by people who enjoy coding, working on lots of projects, and are happy to have many middle-author papers. Oxford have had one of these for years (led by Steve Taylor whom you and I both know). Imperial have one. Nottingham have just set one up. I am sure that there are others.

    • Hi Dov

      I 100% agree on both points – and I hear what you’re saying about rejecting grants on that basis. I fear the horse has bolted though!

      The main opposition to core bioinformatics support groups seems to be funding – I know Nottingham and Newcastle both have them in some form, but they need to recoup their costs from charges to PIs who use the service. Whether this is sustainable, I don’t know.


    • I will have to disagree with the implication that a Biology PI taking on a Bioinformatics PhD student will automatically lead to a disaster and should not be funded. I’m sure the main reason bioinformatics PIs don’t manage lab staff is that they usually don’t have labs, and having some “pet biologists” would be great to get all those predictions tested. As a conflict of interest disclaimer, I am a PhD student with a degree in computer science, and I’m currently working in a lab full of biologists.

      The implication in both your comment and Mick’s original article seems to be that the moment you are the only computational person among biologists, all your time is taken up by ‘routine’ tasks for all the people around you, and if you’re lucky you’ll end up as middle author on some of the papers the “real” scientists publish. The solution you suggest seems to be to institutionalize this by creating a petting zoo to collect all the pet bioinformaticians, so they can at least have some company.

      Of course Mick’s tips make a lot of sense for the lone bioinformatician. I don’t want to claim that working with people who aren’t into your specialisation is unproblematic. Compared to getting my masters degree, I can’t just walk next door and bounce my problem off the heads of fellow computer scientists. Actually, at this university, the comp. sci. people even are on a different campus. But if you keep in contact with the bioinformatics people in other groups, go attend their seminars and convince your PI to send you to bioinformatics conferences, working closely with your prospective users is incredibly rewarding.

      I don’t spend a lot of time with routine tasks that would be handled by a bioinformatics “core facility”. The problems I work on are primary research, and I get to publish first author papers. As a PhD student of course I’m just at the start of my career, but I don’t have the impression that I’m less productive than my fellow students who stayed in the computer science fold.

      I think that no matter what interdisciplinary work you do, you might face a little more challenges than people who travel the well-trodden paths of a given single discipline. But I believe that you also get the opportunity to do relevant work. If the grant that pays for my job had been shot down because there was no bioinformatics PI on the project, I doubt I’d be getting a PhD at the moment. If we refrained from doing things that might go wrong, we wouldn’t be doing science. It’s good to be mindful of potential problems, but a flat-out refusal to fund projects is over the top.


      • I didn’t want to imply that all biologists are bad at training bioinformaticians – I just want to emphasize that training and career development is part of any PhD and post-doc, and it is reasonably common (in my experience) that these two issues are not highest on the agenda when bioinformaticians are employed in lab-based groups.

        The PIs of lab-based groups just need to recognise, I think, that bioinformatics is not a black box, that their students and post-docs might be developing bad practice due to isolation, that they need contact with others of their kind, that they need publications and career progression and that the lab-based PI cannot provide all of those things.

        I am sure some lab-based PIs do recognise these things, but many do not.

        Finally, I am a bioinformatician in charge of a lab – and the only way that works is that I work with an exceptional lab manager who takes care of that side of things, with only strategic input from me. How many lab’s have a senior bioinformatician just like I have a senior lab manager? I’m guessing none.

        • Hi Kai

          Since you disagree with me I think it is polite that I reply too! I too do not imply that /all/ people with an experimental background are incapable of supervising a bioinformatics PI, but this is the case for the majority of situations that I have observed. In general, the two most common failings are:

          (i) PIs who believe that because a particular bioinformatics analysis is so crucial to the science that they are doing that it is appropriate as a bioinformatics PhD or PD project. I have seen many cases where PIs are right about the first part, but wrong about the second: what they are in fact asking the PhD student or PD to do is something that is entirely routine and technical, but not something upon which the student or PD could do novel /bioinformatics/ research, i.e. research that would lead to them gaining first author papers and building a bioinformatics career. The worst I have seen in terms of project is a PI who believed that calculating coefficients of variability for metabolomics data was appropriate for someone’s PhD project; the worst I have seen in terms of personal impact is a bioinformatics PhD student who became suicidally depressed.

          (ii) PIs who do not have the technical knowledge or team in their group to sustain the professional development of their PhD student / PD. This is something Mick has written about at length.

          I have only met a very small number of experimentalists who genuinely have the capacity both to identify appropriate bioinformatics projects and support their staff to deliver them: Francesco Falciani in Liverpool is one such person. I have also met only a very small number of theoreticians who could do the same with experimental work: Julian Lewis of CRUK (and co-author of Molecular Biology of the Cell) is one such person.

  12. I want to disagree (strongly) with Dov’s idea that research councils should not fund bioinformaticians working in biology labs. There are important biological questions that can and should be answered using bioinformatics. This research should be funded as part of the relevant biology, and the bioinformatician doing the research should be immersed in the biology, since they are addressing biological questions. What matters is the biology, and if the bioinformatics is “routine and technical” that’s just fine. Innovative techniques are not a required part of good science. Papers are not routinely rejected because the laboratory methods used have been used before. Rather, one needs to identify and use appropriate techniques to address important questions.
    I see a lot of people wasting their time trying to create innovative software in a space that is already overcrowded. Right on this very blog there was a call for an embargo on short read alignment software.
    We need bioinformaticians who can appreciate the biology and recognize which questions can be addressed with their skills far more than we need most of what might be called new bioinformatics. Put another way, bioinformatics doctorates and postdocs that answer important biological questions making use of their skills will have done more for their career, and for the field, than bioinformatics doctorates and postdocs who produce the 43rd tool for analysis of whatever method is popular this year.
    Having said my peace on that, I’d like to commend Mick for drawing attention to a very important issue. I agree completely that a bioinformatician working on a biological problem must spend a significant amount of time in the company of both biologists and bioinformaticians (or statisticians or computer scientists), that this is a challenge, and that both the PIs and the trainee need to be aware of this need and make it happen. I’ve seen a lot of students and postdocs on both sides, and for every “embedded bioinformatician” who isn’t keeping up wit the best methods and tools there is a computational biologist who is wasting his time, and public funds, creating “innovative bioinformatics” that does not address any real biological question.

    • Hello Steve, I think that the part that is left out is that bioinformatics analysis is indeed taking more and more time (omics experiments are getting more complicated and results need to be compared to external data) but there are no clear guidelines or procedures about for instance shared first authorship. Also the biologist who carried out the experiment still probably did more work on the paper, its just that the bioinformatician also did a lot, but maybe just not enough to warrant 1st authorship. Credit is an all or nothing thing and there is no credit for the 2nd/3rd authorship, even if 2-3 such projects already constitute the same amount of work as a first authorship paper. So in that sense the way merit and authorship are allocated does not really support this team-work way of doing science. In general bioinformatics analysts are career wise better off doing either meta-analysis, web services or tool development (even if it is the 45th alignment tool), even though it might be more profitable for science to be doing analysis of original data. There is also plenty of graduate project / post-doc funding to do bioinformatics analysis. So everyone wins (except the bioinformatician in the long run).

  13. Reblogged this on petbioinformatician and commented:
    Useful advice.

  14. I really enjoyed this article. After I finished my PhD in Applied Math department I took a job being the only mathematician in a department of Leukemia and I’ve been undergoing quite a culture shock. Do you have any advice for PIs in managing “pet bioinformaticians” like me? I’d like to show something to my PI so he can understand how we work and what to expect.

  15. Reblogged this on Pythonic Biologist and commented:
    very useful talk.

  16. Very nice. I wish this wisdom had been widely accepted 8 years ago. Lonely isolated bioinformatics “guy” has been the norm worldwide for a long time. Its the root of file format rot.

  17. Oh boy, that’s so unashamed.
    Reading this opinion i feel my eyes burning.
    If you’re a “pet” (this is unashamed of) bioinformatics and you do your “black magic”, the most of ordinary scientist or PI does not understand because they just push the red button, sorry guy, but you’re asshole.
    You have to look and study our data-set and develop our vision to finally see what the data have bright to you.
    Most ordinary scientist (like biologyst, geneticist, biochemistry) don’t understand what the power inside the data-sets and they really dosen’t care of your results or your hypothesis.
    It’s hard, and you, calling us by “pet”, just prove how much you don’t give a fuck. Sorry man.
    Please, listening what we have to telling you. We have so much to increase and now, the most of our results are bigger and you and other ordinary scientist like you will pass away. “Pet” bioinformaticians will become monsters and you will not like this. Without us, your pappers are nothing.

    Sorry for being so rude. I fight for the bioinformatic field in Brazil since 2010 and today my PI brought this article.

    Nilson AR Coimbra – 24 Years
    PhD Student Bioinformatics
    MSc. BIoinformatics
    Federal University of Minas Gerais.

  18. Venkatesh Chellappa

    22nd March 2017 at 4:43 am

    Ha!! ‘Pet bioinformatician!’ good term.. I was there once upon a time!! a pet bioinformatician caught up in a big project and nowhere to run for help and often took sanctuary in my 4X4 cubicle! I can totally feel the tension.

    Now, 4 years on, any development? Please document your journey and how you got to where you are right now. It will be fun to read and know.

    A big wave to other fellow Pet Bioinformaticians too!
    Venkatesh Chellappa

Leave a Reply

Your email address will not be published.




© 2017 Opiniomics

Theme by Anders NorenUp ↑